Improvement of the literacy of AI in first cycle pre-medical education through student associations: educational intervention | BMC Medical Education

Study design

This study used an almost expression and pretest-post-test control group design to assess the educational impact of a student program led by students. The participants were assigned to an intervention group, which completed the program through AI in the Medicine Association (AIM), a control group, which has not received the intervention. The results were evaluated by knowledge gains in AI and the content linked to pathology.

Participants and recruitment

Participants were BYU undergraduate students pursuing pre-drug or pre-health career paths. The cohort participants were recruited via leaflets, campus bulletins and a stand in person at the Byu student center. Participants in the control group were recruited by awareness-raising in person in major pre-Medés Courses. The inclusion criteria require current registration in pre-drug prices and the interest of health care careers. Students with previous official training on AI beyond the scope of the AIM program were excluded.

Intervention

The AIM program included four weekly educational sessions incorporating active learning principles and multimodal education, as recommended in the literature (10,,11,,12,,13). The program has integrated the concepts of fundamental AI, ethical frameworks, practical commitment with the analysis of histological images and exposure to current research on AI. Specific subjects included:

1. 1 and 1

   Introduction to AI in health care.

2. 2

   Ethical implications and responsible use of AI.

3. 3 and 3

   Data pre -treatment techniques.

4. 4

   Development, training and evaluation of the model.

The teaching material was designed by AIM student officers and examined by a member of the faculty to ensure academic rigor and educational clarity. A key project consisted in identifying cancer tissues and histological characteristics in the images of prostate with whole gliding, promoting practical engagement with AI in pathology (20).

Evaluation design

The main evaluation tool consisted of a structured survey administered both before and after the intervention. He started with six demographic questions / background (for example, an exposure prior to AI, year of graduation, attitudes to AI) followed by ten elements of knowledge divided into two areas:

The formats of articles included multiple and short questions. A full version of the evaluation is available in additional equipment.

Data collection

The intervention and control groups completed the same evaluation before and after a period of four weeks. Evaluations have been carried out in supervised environments or without distraction. Only participants in the cohort who finished the four AIM sessions were eligible for post-test. The witness participants received their post-test four weeks after the pretest, without exposure to the intervention.

To reduce the response bias, the participants were assured that their answers would remain anonymous, were not noted for performance and could omit questions so desired. The identification of the information has been deleted from the scores once collected and used only to contact students if necessary, to connect pre and post-tests to the same student and to collect demographic information which has not been collected in the survey such as the sex of the participant. The participation was voluntary and all the students were informed that they could withdraw at any time. Those who opted for the study were always allowed to fully engage in goal activities. The data was collected on two academic semesters and included 15 participants in the intervention and 16 witnesses.

Notation and notation

The responses were classified by a member of the faculty using a structured section to ensure coherence and minimize biases. Partial credit has been granted for partially correct responses, with specific criteria delimited in the section (see additional equipment).

Statistical analysis

Descriptive statistics (averages, standard deviations, medians, frequencies and percentages) were used to summarize demographic variables and basic characteristics. Comparisons between the groups initially was evaluated using tests of the Chi square (categorical variables) and independent T tests or U mann-Whitney U tests (continuous variables).

To assess the knowledge gains, changes within the group in pre-test and post-test scores were analyzed using apparent T tests (or rank tests signed by Wilcoxon if the data was not distributed normally). Comparisons between post-test scores were analyzed using independent T tests or Mann-Whitney U tests.

The effect sizes were calculated using Cohen D. In addition, an analysis of covariance was carried out to compare post-test scores between groups while controlling pre-test scores and potential confusion factors (for example, prior IA experience, pre-med status).

Exploratory analyzes included correlations of Pearson or Spearman between the prior IA experience and knowledge gains. A multiple regression model was used to identify the predictors of improvement, with the allocation of group, the prior experience of AI and the reference attitudes towards AI have entered as independent variables.

Visualization of data

To visualize distributions and score changes, violin plots have been generated for pre and post-test scores by group. Barre traces with error bars have represented average score changes and standard deviations between AI, pathology and combined areas. Dispersion plots have explored associations between experience or previous attitudes and test performance. All figures have been generated using Python packages marine And matplot.

Ethical considerations

The study received approval from Brigham Young University Institutional Review Board. All participants gave informed consent and data confidentiality was maintained via secure storage and anonymized analysis (see additional equipment).